Body temperature, activity and melatonin profiles in adults with attention-deficit/hyperactivity disorder and delayed sleep: a case control study

Transcription

1 J Sleep Res. (2013) 22, Circadian rhythms and ADHD Body temperature, activity and melatonin profiles in adults with attention-deficit/hyperactivity disorder and delayed sleep: a case control study DENISE BIJLENGA 1, EUS J. W. VAN SOMEREN 2,3, REUT GRUBER 4, TANNETJE I. BRON 1, I. FEMKE KRUITHOF 1, ELISE C. A. SPANBROEK 1 and J. J. SANDRA KOOIJ 1 1 PsyQ Expertise Center Adult ADHD, The Hague, The Netherlands, 2 Department of Sleep & Cognition, Netherlands Institute for Neuroscience, Amsterdam, The Netherlands, 3 Departments of Medical Psychology and Integrated Neurophysiology, Neuroscience Campus Amsterdam, VU University and Medical Center, Amsterdam, The Netherlands and 4 Department of Psychiatry, McGill University, Montreal, QC, Canada Keywords circadian rhythm, core body temperature, delayed sleep phase syndrome, dim-light melatonin onset, skin temperature Correspondence Denise Bijlenga, PhD, Senior Researcher, PsyQ Expertise Center Adult ADHD, Carel Reinierszkade 197, 2593 HR The Hague, The Netherlands. Tel.: +31-(0) ; fax: +31-(0) ; d.bijlenga@psyq.nl Accepted in revised form 17 June 2013; received 14 January 2013 DOI: /jsr SUMMARY Irregular sleep wake patterns and delayed sleep times are common in adults with attention-deficit/hyperactivity disorder, but mechanisms underlying these problems are unknown. The present case control study examined whether circadian abnormalities underlie these sleep problems in a naturalistic home setting. We included 12 medication-naïve patients with attention-deficit/hyperactivity disorder and delayed sleep phase syndrome, and 12 matched healthy controls. We examined associations between sleep/wake rhythm in attention-deficit/hyperactivity disorder and circadian parameters (i.e. salivary melatonin concentrations, core and skin temperatures, and activity patterns) of the patients and controls during five consecutive days and nights. Daily bedtimes were more variable within patients compared with controls (F = 8.19, P < 0.001), but melatonin profiles were equally stable within individuals. Dim-light melatonin onset was about 1.5 h later in the patient group (U = 771, Z = 4.63, P < 0.001). Patients slept about 1 h less on nights before work days compared with controls (F = 11.21, P = 0.002). The interval between dim-light melatonin onset and sleep onset was on average 1 h longer in patients compared with controls (U = 1117, Z = 2.62, P = 0.009). This interval was even longer in patients with extremely late chronotype. Melatonin, activity and body temperatures were delayed to comparable degrees in patients. Overall temperatures were lower in patients than controls. Sleep-onset difficulties correlated with greater distal proximal temperature gradient (DPG; i.e. colder hands, r 2 = 0.32, P = 0.028) in patients. Observed day-to-day bedtime variability of individuals with attention-deficit/hyperactivity disorder and delayed sleep phase syndrome were not reflected in their melatonin profiles. Irregular sleep wake patterns and delayed sleep in individuals with attention-deficit/hyperactivity disorder and delayed sleep phase syndrome are associated with delays and dysregulations of the core and skin temperatures. INTRODUCTION Attention-deficit/hyperactivity disorder (ADHD) is characterized by chronically high levels of inattention, impulsivity and hyperactivity, routinely continues into adulthood, and may result in impairments in, for example, psychological, social, relational, educational and occupational contexts (Biederman, 1998). The estimated international prevalence rate of ADHD in adults is 3.4% (Fayyad et al., 2007). As many as 80% of adults with ADHD have sleep-onset difficulties, 607

2 608 D. Bijlenga et al. which is in most cases associated with a comorbid delayed sleep phase syndrome (DSPS; Van Der Heijden et al., 2005; Van Veen et al., 2010). According to the DSM-IV, DSPS is characterized by chronic late sleep and late rising, and an inability to fall asleep and wake up at earlier times. This can lead to daytime sleepiness, insomnia at night, and impaired functioning in social and occupational contexts (APA, 1994). Adult patients with ADHD with delayed sleep have lower sleep quality, difficulty getting to sleep and difficulty getting up in the morning (Schredl et al., 2007). This may lead to chronic sleep debt in case one has to get up early for school, family life or work (Shirayama et al., 2003). DSPS is already prevalent among children with ADHD, suggesting it may be a chronic condition in this population (Van Der Heijden et al., 2005). We previously reported associations between delayed sleep and hyperactivity/impulsivity symptoms in both patients with ADHD and in the general population (Bijlenga et al., 2013). Moreover, genetic overlaps have been found between ADHD and delayed sleep (Baird et al., 2012). Hyperactive/impulsive symptoms and delayed sleep are thus interrelated, potentially causing difficulties in distinguishing behavioural characteristics that are caused by either of the two. Many patients with ADHD may also be at risk for major health consequences, such as obesity, diabetes, hypertension, cardiovascular disease, immune suppression and even cancer, due to delayed sleep and the resulting chronic sleep debt (Baird et al., 2012). The development of an effective treatment is facilitated by a better understanding of the underlying mechanisms of DSPS in ADHD. The secretion of melatonin by the pineal gland of the brain is delayed in individuals with delayed sleep (Van Someren and Riemersma-Van Der Lek, 2007). The pineal gland is controlled by the hypothalamic suprachiasmatic nucleus (SCN; the body s circadian pacemaker), which in turn receives information about the time of day from the retina of the eye (Guido et al., 2010). The time in the evening at which circulatory melatonin levels reach a certain threshold is called the dim-light melatonin onset (DLMO). Sleepiness and sleep onset occur about 2 h after DLMO (Lewy et al., 1998). Individuals with DSPS have a phase-delayed DLMO, which is thought to contribute to their delays in sleepiness and sleep (Lewy, 2009). Body temperature is also orchestrated by the SCN. Nocturnal release of melatonin is strongly related to a decline of the core body temperature (CBT; Cagnacci et al., 1997). Proximal (chest) skin temperature has a similar, but weaker, relationship to melatonin (Cagnacci et al., 1997). However, skin temperatures from distal regions such as hands and feet have a positive linear relationship to melatonin concentration (Cagnacci et al., 1997). Thus, during wake, melatonin and distal skin temperatures are low while CBT and proximal skin temperatures are high. During sleep, the opposite is true. It is unknown if there is a dysregulation between melatonin release, CBT and skin temperatures in individuals with ADHD and DSPS. Clinical observations suggest that adults with ADHD and DSPS often have large variability in their dayto-day sleep/wake rhythm. They tend to sleep at highly variable times, ranging from normal to extremely late. We investigated the intra-individual variability of the DLMO, and related our findings to the activity profiles of the patients and to healthy matched controls. We investigated whether the relationships between the CBT, skin temperature and melatonin profiles had the same phase delay as the activity profiles in the patients. The results of this study will provide insight into the physical and behavioural characteristics associated with DSPS in adults with ADHD, and contribute to methods to prevent chronic sleep deprivation and guard against its associated health risks. MATERIALS AND METHODS Participants Patients were recruited from the PsyQ outpatient clinic (Program Adult ADHD, The Hague, the Netherlands). Diagnostic assessment of ADHD was performed by a trained psychologist using the Diagnostic Interview for ADHD in adults (DIVA 2.0), which is based on DSM-IV criteria for ADHD (Kooij and Francken, 2010). Also common psychiatric and physical comorbidities were systematically assessed. DSPS was defined as the inability to fall asleep at a preferred time before 23:30 h within the past 6 months, along with reports of a negative impact of the delayed sleep on the patient s professional and/or social life. We calculated the mid-sleep on free days corrected for sleep debt on work days (MSFsc) to measure chronotype using the Munich Chronotype Questionnaire (MCTQ; Roenneberg et al., 2007). We recruited healthy controls by invitations. Controls were matched to enrolled patients in terms of sex and age (5 years). Each patient and control match was enrolled within the same season to avoid effects of day length and daylight intensity. All participants received compensation for travel expenses. Controls received a small compensation for study participation. The Dutch Medical Ethical Committee for Mental Health Research (METiGG) approved this study in May 2009 (NL , Utrecht, the Netherlands). Inclusion criteria were age between 18 and 55 years and fluency in the Dutch language. Additional inclusion criteria for patients were clinical diagnoses of both ADHD and DSPS. ADHD symptoms, delayed sleep and use of any psychiatric medication were exclusion criteria for the control group. Exclusion criteria for both groups were (a history of) psychotic disorder, anxiety disorder, depression, high alcohol intake (men >21 and women >15 units week 1 ), use of nonprescribed drugs within a month prior to study participation, use of medications that affect sleep, suspected cognitive dysfunction, mental retardation, conducting shift work, having crossed >2 time zones within 2 weeks prior to study participation, and/or having had bright light therapy within a month prior to study participation.

3 Temperature, activity and melatonin in sleep delay and ADHD 609 Procedures Eligible participants were assessed face-to-face by one of the researchers (D. B. or T. I. B.). Both written and oral information about the study were provided, and inclusion and exclusion criteria were systematically rechecked. Participants signed informed consent and filled out the Sleep Hygiene Questionnaire. During the 5-day study period that always started on a Saturday, participants were subjected to actigraphy, CBT, skin temperature and melatonin measurements. Participants filled out a sleep log. The measurements were performed in the participant s naturalistic home setting. participants to use dimmed light during saliva sampling, which has a minimal effect on circulatory melatonin levels (Pullman et al., 2012; Wahnschaffe et al., 2013). Melatonin concentrations were determined by radioimmunoassay at the Gelderse Vallei Hospital, Ede (the Netherlands). Details on the assaying method can be found elsewhere (Zeitzer et al., 2000). Values of 0 pg ml 1 were assigned to concentrations below the sensitivity threshold of 0.5 pg ml 1. A melatonin concentration of 3.0 pg ml 1 was used as the DLMO (Benloucif et al., 2008). We calculated the time of DLMO by linear interpolation of the melatonin concentrations. Any DLMO before 21:00 h was set at 21:00 h; any DLMO after 03:00 h was set at 03:00 h. Sleep hygiene The Sleep Hygiene Questionnaire consisted of 22 items, scored on a five-point Likert scale ranging from 1 never to 5 (almost) always. The questionnaire was based on the Adolescent Sleep Hygiene Scale (LeBourgeois et al., 2004), and on findings on the relationship between sleep hygiene and sleep quality by Harvey (2000) and Jefferson et al. (2005). Total score ranged from 22 to 110, a higher score indicating worse sleep hygiene. Sleep logs The participants kept a daily sleep log. In the evening they reported their daytime activities, caffeinated drinks and cigarette consumption, and the times they did not wear their research devices (e.g. whilst taking a shower). On the following morning they reported their bedtime, sleep-onset latency (SOL) duration, number and duration of wake bouts, and time of wake-up and rise. Actigraphy Activity was continuously monitored using ActiWatch â -L (AW64 series; Mini Mitter, Bend, OR, USA). Actigraphy is widely used to estimate sleep parameters, and has been validated against polysomnography with very high agreement rates (Kushida et al., 2001). An ActiWatch is a wristwatch-like device that monitors accelerations generated by movements and is minimally invasive. The participants wore the ActiWatch on the wrist of their non-dominant hand continuously during the full 5-day study period. We used 30-s epochs. Bedtime and wake-up times from sleep logs were used as the start and end for the sleep parameters. Salivary melatonin Saliva was collected with Salivette â cotton swabs (Sarstedt, Nümbrecht, Germany). The sampling was every evening of the five study days, every hour from 21:00 to 03:00 h. Because light suppresses melatonin, we instructed the Core body temperature Core body temperature was monitored continuously using VitalSense â (Philips Respironics, Mini Mitter), a wireless integrated physiological monitor. The VitalSense consists of a data logger with a resolution of 0.01 C, which receives radio signals from the temperature-sensitive ingestible capsules with a sensing range between 25 and 50 C (0.10 C). VitalSense generates a valid index of the CBT (Darwent et al., 2011). We used 1-min epochs. Because within the first 5 h of capsule transit systematic errors have been noted (Darwent et al., 2011), we have erased CBT data of this period. At excretion of a capsule h after intake, a new capsule was ingested. The logger s reception range is 1 m; participants therefore had to adjust the logger to their belt or carry it in their handbag during the daytime, and to place the logger in or by their bed at night. Skin temperature Skin temperature was continuously monitored using three Thermochron ibuttons â (DS1922L-F5; Maxim Integrated Products, San Jose, CA, USA), which are wireless peripheral thermometry devices measuring temperatures from 10 to 65 C (0.5 C). An ibutton consists of a 3 V Lithium battery, semiconductor temperature sensor, computer chip integrating a 1-Wire transmitter/receiver, clock, thermal history log and memory storage, which is all enclosed in a round, 16-mm stainless-steel can. ibutton is a reliable and valid method to measure skin temperature in humans (Hasselberg et al., 2011; Sarabia et al., 2008). We used 30-s epochs. Proximal skin temperature was measured on the sub-clavicular site with an ibutton attached to the skin using sports tape. Distal skin temperatures were measured on the dorsal side of the base of the middle finger and on the dorsal side of the wrist, both on the non-dominant hand. The mean of these sites was the distal skin temperature. The ibuttons were held in place using Velcro finger- and wristbands. The distal proximal temperature gradient (DPG), the distal skin temperature minus the proximal skin temperature, has been shown to reflect sleep propensity accurately (Hasselberg et al., 2011).

4 610 D. Bijlenga et al. Sample size We used a one-sided t-test with a = 0.05, 1 b = 0.80 and an effect size of The means and standard deviations (SDs) of the intra-individual variability in DLMO were adapted from previous results (Van Veen et al., 2010): the mean of the patient group was 79 min, and 25 min in controls. The SD of intra-individual variability was estimated at 40 min. A minimal sample size of 11 participants per group was needed. We added one participant per group to correct for missing data. We used a total sample size of N =24 participants (12 patients and 12 controls). Analyses General characteristics and questionnaire outcomes were analysed using t-tests for continuous normal distributed data and Mann Whitney s U-test for non-parametric data. v 2 -tests were used for categorical data. The intra-individual variability of DLMO was defined as the variance of DLMO per participant over 5 days, and was compared between groups by one-way ANOVA. Pearson s correlations were used to separately examine the association between DLMO and sleep time. The slope of the melatonin curve at DLMO and activity levels in the 10-min intervals in which DLMOs occurred were examined. We correlated the mean activity per 10-min interval to mean CBT per interval using Pearson s correlation (r). We tested between-group differences of activity levels and temperatures using one-way ANOVAs. The 10-min time intervals of the minimum (T min ) and maximum (T max ) temperatures were compared between groups by oneway ANOVAs. We used mean temperatures of 20-min intervals (10 min before and 10 min after every hour) and correlated them with melatonin concentrations that were sampled on the hour using Pearson s correlation (r). SOL was correlated to distal skin temperature and DPG for the 10-min interval in which the individual reported bedtime. SPSS version 20.0 for Windows (SPSS, Chicago, IL, USA) was used, and P-values 0.05 indicated statistical significance. RESULTS Between July 2010 and September 2011, we enrolled 12 patients and 12 matched healthy controls. Half of the participants were male, and all were between 21 and 55 years old (mean 32.4 years, SD 10.1). General characteristics are shown in Table 1. Table 2 compares times of mid-sleep, sleep hygiene, subjective sleep logs, objective actigraphy and salivary melatonin characteristics between patients and controls. Patients reported a later mid-sleep than controls with a mean difference of almost 2 h (U =8, Z = 3.70, P < 0.001). Patients had worse sleep hygiene (U = 33, Z = 2.26, P = 0.024), scoring worse on the following items: going to bed about the same time every night ; being hungry at bedtime ; having telephone conversations while in bed at Table 1 General characteristics of the patient and control groups, N =24 Patients, N =2 Controls, N = 12 Value P General Males: N (%) 6 (50) 6 (50) v 2 = Age (years): 32.5 (9.9) 32.4 (10.8) t = mean (SD) Body mass index 23.4 (3.9) 22.8 (2.0) t = (kg m 2 ): mean (SD) Work days 2.7 (2.1) 3.5 (1.8) t = per week: mean (SD) Smokes: N (%) 4 (33.3) 3 (25.0) v 2 = If yes, 9.1 (7.1) 4.0 (6.7) t = how many cigarettes per day: mean (SD)* Caffeine: 2.9 (1.4) 2.5 (2.3) t = mean units per day (SD)* Alcohol: 0.6 (0.8) 0.6 (0.9) t = mean units per day (SD)* Sports: mean duration per day (in min) (SD)* 0:26 (0:27) 0:36 (0:27) t = *Data derived from subjective sleep logs. night ; watching TV while in bed at night ; and getting up about the same time every morning. Time of lights out was almost 2.5 h later in patients than controls (U = 4, Z = 3.93, P < 0.001). Time of falling asleep was about 2 h and 20 min later in patients than controls (sleep logs: U =2, Z = 4.04, P < 0.001; actigraphy: F = 7.36, P = 0.013). There was larger day-to-day intra-individual variability in the time of falling asleep in patients compared with controls (F = 8.19, P < 0.001). Sleep duration was about 1 h shorter for patients than controls on nights prior to work days (F = 11.21, P = 0.002). Patients reached DLMO about 1.5 h later than controls (U = 771, Z = 4.63, P < 0.001). The day-to-day intra-individual variability of time of DLMO was comparable (F = 0.16, P = 0.695). Time of DLMO did not differ between nights before work days and nights before free days for patients (F = 0.11, P = 0.743) or controls (F = 1.21, P = 0.276; data not shown). The time between DLMO and falling asleep was 1 h longer in patients (U = 1117, Z = 2.62, P = 0.009). The duration between time of DLMO and falling asleep was smaller, and not significantly different, on nights before work days (U = 241, Z = 0.65, P = 0.515). On nights before free days this was near-significantly different (U = 316, Z = 1.95, P = 0.052). Fig. 1 shows correlations between time of DLMO and sleep onset. We investigated the time of sleep onset using a two-way mixed model to control for multiple observations within subjects. The main effect Group and interaction effect

5 Temperature, activity and melatonin in sleep delay and ADHD 611 Table 2 Mid-sleep times and sleep hygiene, subjective and objective sleep measures, and salivary melatonin within the 5-day study period: comparisons between patients and controls using Mann Whitney s U-test, ANOVAs or repeated-measures mixed models, N = 24 Patients, N =12 Mean (SD) Controls, N =12 Mean (SD) Value P MCTQ questionnaire MSFsc 06:08 h (1:15) 04:15 h (0:37) U(Z) =8( 3.70) <0.001* Sleep hygiene questionnaire Total score (range ; higher 52.8 (6.5) 46.9 (6.5) U(Z) = 33( 2.26) 0.024* score = worse sleep hygiene) Self-reported sleep characteristics from sleep log Time of lights out to sleep 02:31 h (1:12) 00:08 h (0:38) U(Z) =4( 3.93) <0.001* Time of falling asleep 02:52 h (1:11) 00:28 h (0:35) U(Z) =2( 4.04) <0.001* Intra-individual variability in time of falling asleep 1:05 (0:34) 0:42 (0:24) F = 8.19 <0.001* Objective sleep characteristics from actigraphy Time of falling asleep 02:36 h (1:08) 00:16 h (0:41) F = * Number of wake bouts each night 35.6 (12.5) 36.0 (7.7) U(Z) = 529 ( 1.85) Sleep duration all nights 6:13 (1:28) 6:24 (1:23) F = Sleep duration on nights before work days 4:54 (0:51) 6:02 (1:18) F = * Sleep duration on nights before free days 6:48 (1:17) 7:01 (1:19) F = Sleep efficiency in % 81.8 (9.2) 77.3 (1.3) U(Z) = 1249( 2.77) 0.006* SOL in minutes 13.3 (11.3) 17.3 (14.5) U(Z) = 683 ( 0.75) Salivary melatonin characteristics Time of DLMO 00:05 h (1:35) 22:38 h (1:03) U(Z) = 771( 4.63) <0.001* Intra-individual variability in time of DLMO 0:35 (0:23) 0:38 (0:22) F = Salivary melatonin and sleep log Time between DLMO and falling asleep all nights 2:28 (2:00) 1:28 (1:25) U(Z) = 1117( 2.62) 0.009* Time between DLMO and falling asleep on 1:48 (1:42) 1:18 (1:26) U(Z) = 241( 0.65) nights before work days (N = 19) Time between DLMO and falling asleep on nights before free days 2:44 (2:04) 1:44 (1:22) U(Z) = 316 ( 1.95) DLMO, dim-light melatonin onset, as assessed using a salivary melatonin concentration threshold of 3.0 pg ml 1 ; MCTQ, Munich Chronotype Questionnaire; MSF sc, mid-sleep on free days, corrected for sleep debt on work days. *Significant with a < Group 9 Time of DLMO were non-significant, but main effect Time of DLMO was near-significant (b = 0.26, SE = 0.14, P = 0.065). Correlation for the total group after controlling for repeated measures within subjects was r = 0.23 (P = 0.014), for the patient group r = 0.27 (P = 0.045), and control group r = 0.14 (P = 0.321). Exclusion of three outliers (in the top of Fig. 1) from the patient group did not alter correlations. The interpolated lines show that the later the DLMO, the higher the between-group difference, with an approximately 1.5-h difference for those with DLMOs at 21:00 h increasing to an approximately 3-h difference for those with DLMOs at 02:00 h. Fig. 2 shows the relationship between mean activity level and the mean melatonin slope per group. Both melatonin and activity profiles were delayed in patients. The melatonin slopes within 1 h after DLMO were similar (patients M = 3.11 pg ml 1, SD = 2.50 versus controls M = 3.53 pg ml 1, SD = 2.27; t = 0.86, P = 0.391). Mean activity levels in the 20-min interval around the time of DLMO were also comparable between groups (patients M = , SD = versus controls M = , SD = 93.75; F = 1.56, P = 0.213). Figure 1. Correlations between time of dim-light melatonin onset (DLMO) and sleep onset in the patient (solid symbols and solid line; N = 12) and control (open symbols and dotted line; N = 12) groups.

6 612 D. Bijlenga et al. Figure 2. The relationship between the mean actigraphy-measured activity counts (grey line) and the mean melatonin buildup (black line) per group, with 95% confidence intervals. The dotted horizontal line indicates the dim-light melatonin onset (DLMO), as assessed by a salivary melatonin threshold of 3 pg ml 1. Fig. 3a shows the mean activity levels during the total study period for both groups. The activity levels of patients were delayed by 2 3 h compared with controls. Mean activity levels at bedtimes and wake-up times did not differ between groups. The two-way repeated measures main effects mixed model showed no main effect of group on total activity (b = 4.09, SE = 11.89, P = 0.734). Mean activity levels during sleep periods were higher in patients (patients M = 22.08, SD = versus controls M = 17.59, SD = 69.14; F = 92.10, P < 0.001). Activity levels were comparable during waking periods (patients M = , SD = versus controls M = , SD = ; F = 0.18, P = 0.671). Activity and CBT highly correlated in both patients (r = 0.939, P < 0.001) and controls (r = 0.929, P < 0.001). Means and SEs of activity levels, CBT, distal and proximal skin temperatures, and DPG per group are shown in Fig. 3 a e. As shown in Table 3, the CBT was slightly, though significantly, lower in patients compared with controls. The mean proximal skin temperature was 0.11 C higher (P = 0.001), the mean distal skin temperature was 0.47 C lower (P = 0.014), and the mean DPG was 0.56 C lower (P = 0.001) in patients. The T min and T max temperatures did not differ significantly between groups. Correlations between melatonin concentrations and the measured temperatures are shown in Table 4 (upper panel). In patients, correlations between melatonin concentration and distal temperature (r 2 = 0.15, P = 0.012) and DPG (r 2 = 0.14, P = 0.020) were weak but significant. There was no significant correlation in the control group. As shown in Table 4 (lower panel), SOL as reported in the sleep logs was correlated with each of the temperature measures. We used sleep logs instead of Actigraphy because of statistical power issues. There were significant correlations in the patients between SOL and CBT (r 2 = 0.36, P = 0.028), SOL and DPG (r 2 = 0.32, P = 0.028), and a near-significant correlation between SOL and distal temperature (r 2 = 0.29, P = 0.051). There were no significant correlations in the controls. DISCUSSION Symptoms of hyperactivity/impulsivity and delayed sleep are interrelated. As a consequence, there is a high prevalence of DSPS in adults with ADHD. We investigated the relationship between clinically observed large day-to-day variability of bedtimes among adults with ADHD and DSPS and their melatonin, temperature and activity profiles, and compared these with healthy controls with normal chronotypes. The main findings are that: (i) patients sleep an hour less than controls on nights prior to work days but not prior to free days; (ii) individuals with ADHD and DSPS have large intraindividual day-to-day variability in bedtimes, but this is not caused by day-to-day variability of time of melatonin release onset; (iii) the duration between DLMO and time of falling asleep was about 1 h longer in the patients versus controls,

7 Temperature, activity and melatonin in sleep delay and ADHD 613 (a) (b) (c) (d) (e) Figure 3. Mean waveforms and SEs of (a) activity levels, (b) core body temperature (CBT), (c) distal and (d) proximal skin temperatures, and the (e) distal proximal temperature gradient (DPG) of patients (black) and controls (grey). larger on nights prior to free days, and larger for patients with extremely late DLMOs; (iv) melatonin profiles were later but not different between groups; (v) the temperature, melatonin and activity parameters were all delayed to the same magnitude in the patients, but the relationships between the parameters were equal between groups. As previously noted,

8 614 D. Bijlenga et al. Table 3 Differences between groups in means, T min and T max of the CBT, the distal and proximal temperatures, and the DPG, N =24 Patients, N =12 Mean (SD) Controls, N =12 Mean (SD) Value P CBT (0.30) (0.29) F = T min for CBT (0.33) (0.45) t = T max for CBT (0.31) (0.38) t = Proximal skin temperature (0.19) (0.35) F = * T min for proximal skin temperature (0.72) (1.09) t = T max for proximal skin temperature (0.59) (0.71) t = Distal skin temperature (1.57) (1.66) F = * T min for distal skin temperature (3.84) (2.78) t = T max for distal skin temperature (1.19) (0.87) t = DPG 3.00 (1.46) 2.44 (1.36) F = * T min for DPG 8.00 (3.79) 7.21 (2.73) t = T max for DPG 1.22 (1.26) 1.06 (1.75) t = CBT, core body temperature; DPG, distal proximal temperature gradient. *Significant with a < Table 4 Pearson s correlations between melatonin concentration, SOL and the temperature measurements ( C) per group, N = 24 Patients, N =12 Correlation (r 2 ) P Controls, N =12 Correlation (r 2 ) P Correlation between melatonin concentrations (pg ml 1 ) and: CBT Proximal skin temperature Distal skin temperature * DPG * Correlation between SOL (min) and: CBT * Proximal skin temperature Distal skin temperature DPG * CBT, core body temperature; DPG, distal proximal temperature gradient; SOL, sleep-onset latency. *Significant with a < these patients may be at risk for major chronic health conditions, such as mood disorders, obesity, cardiovascular disease, diabetes and metabolic syndrome because of chronic sleep debt (Copinschi, 2005; Lewy, 2009). The patients did not sleep enough during the work week, and had more movements in their sleep than controls. There was a large difference in sleep duration between nights before work days and free days in patients, which was consistent with earlier findings by Roenneberg et al. (2007), who reported larger differences in bedtimes in people with later chronotypes (Roenneberg et al., 2007). Also, the nocturnal melatonin signal has been shown to be shorter in children with ADHD as compared with healthy children (Novakova et al., 2011). While we did not assess melatonin throughout the night, we do not know the duration of our patients melatonin signal. They slept shorter during the work week, which may be due to either a shorter melatonin signal in general or to a later melatonin onset in the evening in combination with early morning obligations during the work week. The correlations between DLMO and sleep onset were weak but significant, and comparable to those found in the Benloucif et al. (2005) study. The patients had overall worse sleep hygiene, which may explain why their bedtimes were not solely directed by their melatonin profiles. In particular, outside of our research setting, individuals with ADHD more often reported being hungry at bedtime, having telephone conversations in bed, and watching TV while in bed. Light at night (e.g. by the TV or computer screen) is known to suppress melatonin levels (Figueiro et al., 2011). However, in our study the participants were instructed to use dimmed light after 21:00 h. Therefore, we have indications that the delayed melatonin profiles in patients were not delayed by bright light in the evening during the study period. Dim-light melatonin onset, activity and temperature parameters were all equally delayed in the patients, but the relationship between the parameters was comparable between groups. We found lower overall core and distal skin temperatures in patients compared with controls,

9 Temperature, activity and melatonin in sleep delay and ADHD 615 suggesting a potential dysfunction of thermostasis in the patients. Furthermore, having colder hands relative to chest skin temperature was related to longer SOLs among patients. Previous studies showed that patients with cold extremities (feet) and sleep-onset difficulties may suffer from suboptimal distal vasodilatation, which renders them unable to sleep (Krauchi et al., 2008; Lack et al., 2008). This may also be the case for the ADHD group in our study who had cold extremities and long SOLs. Due to our choice of patient population we do not know if these results can be attributed to DSPS in general, or specifically to DSPS in ADHD. This work has some limitations. First, our study was performed in the participants home environment, making it difficult to correct for factors such as late-night light exposure (e.g. from a computer screen) and changes in room temperature, both of which may influence salivary melatonin concentrations (Figueiro et al., 2011; Van Someren and Riemersma-Van Der Lek, 2007). However, we chose the naturalistic setting in order to obtain results that can be generalized to the usual circumstances in the patient s environment. We standardized our procedures as much as possible by instructing the participants to use dimmed light during sampling, and we enrolled patient control matches within the same season to avoid large differences in daylight exposure. Second, our estimation of the time of DLMO may have been unreliable in participants who had low overall melatonin levels. This caused DLMO times to be late in some cases, even though these controls did not have any sleep problems. The mean times of DLMO were about 1 h later than those found in our previous study (Van Veen et al., 2010), which could be due to the fact that many of our patients and some of the controls did not reach the DLMO threshold before 03:00 h. In the previous study, the DLMOs of such participants were set at 01:00 h (the time of the last measurement), whereas they were set to 03:00 h in the present study. Lastly, we found unexpected short sleep in the control group, which may be due to the frequent waking times needed for melatonin sampling. This was particularly the case in controls, who went to bed at a regular time and had to set their alarms every hour between their bedtime and 03:00 h. The patients, however, were often still awake at the sampling times. We therefore presume that, especially for the controls, regular sleep lengths are longer than we measured in the present study. Our study has shown that the melatonin, activity and temperature parameters were out of phase, and temperature was somewhat dysregulated in the patients with DSPS and ADHD as compared with controls with normal chronotype. However, the chaotic sleep/wake rhythm in patients is not caused by variable melatonin profiles. Additional research should examine whether better sleep hygiene and more optimal distal vasodilatation could improve the sleep of patients with DSPS and ADHD, potentially lowering some major health risks in this population. ACKNOWLEDGEMENTS The authors thank all study participants and the staff of PsyQ Program Adult ADHD for screening and referring potential study participants. We thank Inge van Kasteren for her administrative support, Dr Anet Klaassens for her preparatory work, and Dr Marijke Gordijn and Prof. Alfred Lewy for their helpful advice. AUTHOR CONTRIBUTIONS EJWS, RG, MFK, ECAS and JJSK designed the study. DB and TIB conducted the inclusion and entered data. DB performed statistical analyses and reported results. All authors reviewed the manuscript. CONFLICT OF INTEREST This study has not been financially supported by any external provider. JJSK has been a speaker for Janssen B.V. and Lilly B.V., and has received unrestricted research grants from Janssen B.V. and Shire B.V. for other studies. DB, EJWS, RB, TIB, MFK and ECAS declare to have no conflicts of interest. REFERENCES APA. Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), 4th edn. American Psychiatric Association, Washington DC, Baird, A. L., Coogan, A. N., Siddiqui, A., Donev, R. M. and Thome, J. Adult attention-deficit hyperactivity disorder is associated with alterations in circadian rhythms at the behavioural, endocrine and molecular levels. Mol. Psychiatry, 2012, 17: Benloucif, S., Guico, M. J., Reid, K. J., Wolfe, L. F., L hermite- Baleriaux, M. and Zee, P. C. Stability of melatonin and temperature as circadian phase markers and their relation to sleep times in humans. J. Biol. Rhythms, 2005, 20: Benloucif, S., Burgess, H. J., Klerman, E. B. et al. Measuring melatonin in humans. J. Clin. Sleep Med., 2008, 4: Biederman, J. Attention-deficit/hyperactivity disorder: a life-span perspective. J. Clin. Psychiatry, 1998, 59: Bijlenga, D., Van Der Heijden, K. B., Breuk, M. et al. Associations between sleep characteristics, seasonal depressive symptoms, lifestyle, and ADHD symptoms in adults. J. Atten. Disord., 2013, 17: Cagnacci, A., Krauchi, K., Wirz-Justice, A. and Volpe, A. Homeostatic versus circadian effects of melatonin on core body temperature in humans. J. Biol. Rhythms, 1997, 12: Copinschi, G. Metabolic and endocrine effects of sleep deprivation. Essent. Psychopharmacol., 2005, 6: Darwent, D., Zhou, X., Van Den Heuvel, C., Sargent, C. and Roach, G. D. The validity of temperature-sensitive ingestible capsules for measuring core body temperature in laboratory protocols. Chronobiol. Int., 2011, 28: Fayyad, J., De Graaf, R., Kessler, R. et al. Cross-national prevalence and correlates of adult attention-deficit hyperactivity disorder. Br. J. Psychiatry, 2007, 190: Figueiro, M. G., Wood, B., Plitnick, B. and Rea, M. S. The impact of light from computer monitors on melatonin levels in college students. Neuro Endocrinol. Lett., 2011, 32:

10 616 D. Bijlenga et al. Guido, M. E., Garbarino-Pico, E., Contin, M. A. et al. Inner retinal circadian clocks and non-visual photoreceptors: novel players in the circadian system. Prog. Neurobiol., 2010, 92: Harvey, A. G. Sleep hygiene and sleep-onset insomnia. J. Nerv. Ment. Dis., 2000, 188: Hasselberg, M. J., Mcmahon, J. and Parker, K. The validity, reliability, and utility of the ibutton(r) for measurement of body temperature circadian rhythms in sleep/wake research. Sleep Med., 2013, 14: Jefferson, C. D., Drake, C. L., Scofield, H. M. et al. Sleep hygiene practices in a population-based sample of insomniacs. Sleep, 2005, 28: Kooij, J. J. and Francken, M. H. Diagnostic Interview for ADHD in Adults Version 2.0 (DIVA 2.0) [Dutch: Diagnostisch Interview voor ADHD bij volwassenen]. DIVA Foundation, The Hague, the Netherlands, Available online at: Krauchi, K., Gasio, P. F., Vollenweider, S. et al. Cold extremities and difficulties initiating sleep: evidence of co-morbidity from a random sample of a Swiss urban population. J. Sleep Res., 2008, 17: Kushida, C. A., Chang, A., Gadkary, C., Guilleminault, C., Carrillo, O. and Dement, W. C. Comparison of actigraphic, polysomnographic, and subjective assessment of sleep parameters in sleep-disordered patients. Sleep Med., 2001, 2: Lack, L. C., Gradisar, M., Van Someren, E. J., Wright, H. R. and Lushington, K. The relationship between insomnia and body temperatures. Sleep Med. Rev., 2008, 12: LeBourgeois, M. K., Avis, K., Mixon, M., Olmi, J. and Harsh, J. Snoring, sleep quality, and sleepiness across attention-deficit/ hyperactivity disorder subtypes. Sleep, 2004, 27: Lewy, A. J. Circadian misalignment in mood disturbances. Curr. Psychiatry Rep., 2009, 11: Lewy, A. J., Bauer, V. K., Ahmed, S. et al. The human phase response curve (PRC) to melatonin is about 12 hours out of phase with the PRC to light. Chronobiol. Int., 1998, 15: Novakova, M., Paclt, I., Ptacek, R., Kuzelova, H., Hajek, I. and Sumova, A. Salivary melatonin rhythm as a marker of the circadian system in healthy children and those with attention-deficit/hyperactivity disorder. Chronobiol. Int., 2011, 28: Pullman, R. E., Roepke, S. E. and Duffy, J. F. Laboratory validation of an in-home method for assessing circadian phase using dim light melatonin onset (DLMO). Sleep Med., 2012, 13: Roenneberg, T., Kuehnle, T., Juda, M. et al. Epidemiology of the human circadian clock. Sleep Med. Rev., 2007, 11: Sarabia, J. A., Rol, M. A., Mendiola, P. and Madrid, J. A. Circadian rhythm of wrist temperature in normal-living subjects: a candidate of new index of the circadian system. Physiol. Behav., 2008, 95: Schredl, M., Alm, B. and Sobanski, E. Sleep quality in adult patients with attention deficit hyperactivity disorder (ADHD). Eur. Arch. Psychiatry Clin. Neurosci., 2007, 257: Shirayama, M., Shirayama, Y., Iida, H. et al. The psychological aspects of patients with delayed sleep phase syndrome (DSPS). Sleep Med., 2003, 4: Van Der Heijden, K. B., Smits, M. G., Van Someren, E. J. and Gunning, W. B. Idiopathic chronic sleep onset insomnia in attention-deficit/hyperactivity disorder: a circadian rhythm sleep disorder. Chronobiol. Int., 2005, 22: Van Someren, E. J. and Riemersma-Van Der Lek, R. F. Live to the rhythm, slave to the rhythm. Sleep Med. Rev., 2007, 11: Van Veen, M. M., Kooij, J. J., Boonstra, A. M., Gordijn, M. C. and Van Someren, E. J. Delayed circadian rhythm in adults with attention-deficit/hyperactivity disorder and chronic sleep-onset insomnia. Biol. Psychiatry, 2010, 67: Wahnschaffe, A., Haedel, S., Rodenbeck, A. et al. Out of the lab and into the bathroom: evening short-term exposure to conventional light suppresses melatonin and increases alertness perception. Int. J. Mol. Sci., 2013, 14: Zeitzer, J. M., Dijk, D. J., Kronauer, R., Brown, E. and Czeisler, C. Sensitivity of the human circadian pacemaker to nocturnal light: melatonin phase resetting and suppression. J. Physiol., 2000, 526 (Pt 3):